[turbofan] Introduce CheckHole and CheckHoleNaN operators.

[turbofan] Introduce CheckHole and CheckHoleNaN operators.

These simplified operators are used to perform the hole checks when
loading elements from a holey array. Depending on the CheckHoleMode,
they either return the hole as undefined or some NaN, or deoptimize
if the value is the hole or the hole NaN.

R=jarin@chromium.org

Committed: https://crrev.com/502dd40c3307aaa60ee24251f13c9d36e11923a8
Cr-Commit-Position: refs/heads/master@{#37001}

[Benedikt Meurer, 2016-06-15 10:56:39]

Hey Jaro,

Here's the hole check thing.
Please take a look.

Thanks,
Benedikt

[Benedikt Meurer, 2016-06-15 11:00:25]

The CQ bit was checked by bmeurer@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-06-15 11:00:27]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2066223002/1

[commit-bot: I haz the power, 2016-06-15 11:15:31]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-06-15 11:15:32]

Dry run: Try jobs failed on following builders:
  v8_win_nosnap_shared_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng...)
  v8_win_nosnap_shared_rel_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng...)

[Jarin, 2016-06-15 11:18:43]

lgtm.

https://codereview.chromium.org/2066223002/diff/1/src/compiler/effect-control...
File src/compiler/effect-control-linearizer.cc (right):

https://codereview.chromium.org/2066223002/diff/1/src/compiler/effect-control...
src/compiler/effect-control-linearizer.cc:1312: case
CheckHoleMode::kAllowReturnHole:
This is confusing - it does not return hole, it turns hole into undefined. Could
we have a name that somehow implies that (Like kConvertHoleToUndefined or some
such)?

https://codereview.chromium.org/2066223002/diff/1/src/compiler/effect-control...
src/compiler/effect-control-linearizer.cc:1318: control = effect =
graph()->NewNode(common()->DeoptimizeIf(), check,
kDeoptimizeIfHole?

https://codereview.chromium.org/2066223002/diff/1/src/compiler/opcodes.h
File src/compiler/opcodes.h (right):

https://codereview.chromium.org/2066223002/diff/1/src/compiler/opcodes.h#newc...
src/compiler/opcodes.h:228: V(CheckHoleNaN)                  \
How about somehow making more explicit the representation? (The former checks
tagged hole, the latter float64 hole nan, no?)

CheckTaggedHole, CheckFloat64HoleNaN?

[Benedikt Meurer, 2016-06-15 11:39:09]

https://codereview.chromium.org/2066223002/diff/1/src/compiler/effect-control...
File src/compiler/effect-control-linearizer.cc (right):

https://codereview.chromium.org/2066223002/diff/1/src/compiler/effect-control...
src/compiler/effect-control-linearizer.cc:1312: case
CheckHoleMode::kAllowReturnHole:
On 2016/06/15 11:18:43, Jarin wrote:
> This is confusing - it does not return hole, it turns hole into undefined.
Could
> we have a name that somehow implies that (Like kConvertHoleToUndefined or some
> such)?

Done.

https://codereview.chromium.org/2066223002/diff/1/src/compiler/effect-control...
src/compiler/effect-control-linearizer.cc:1318: control = effect =
graph()->NewNode(common()->DeoptimizeIf(), check,
I kinda prefer the never-return-hole thing here.

https://codereview.chromium.org/2066223002/diff/1/src/compiler/opcodes.h
File src/compiler/opcodes.h (right):

https://codereview.chromium.org/2066223002/diff/1/src/compiler/opcodes.h#newc...
src/compiler/opcodes.h:228: V(CheckHoleNaN)                  \
On 2016/06/15 11:18:43, Jarin wrote:
> How about somehow making more explicit the representation? (The former checks
> tagged hole, the latter float64 hole nan, no?)
> 
> CheckTaggedHole, CheckFloat64HoleNaN?

Done.

[Benedikt Meurer, 2016-06-15 11:50:18]

The CQ bit was checked by bmeurer@chromium.org

[Benedikt Meurer, 2016-06-15 11:50:19]

The patchset sent to the CQ was uploaded after l-g-t-m from jarin@chromium.org
Link to the patchset: https://codereview.chromium.org/2066223002/#ps20001
(title: "Address comments. Blacklist test")

[commit-bot: I haz the power, 2016-06-15 11:50:27]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2066223002/20001

[commit-bot: I haz the power, 2016-06-15 12:24:16]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-06-15 12:25:38]

Description was changed from

==========
[turbofan] Introduce CheckHole and CheckHoleNaN operators.

These simplified operators are used to perform the hole checks when
loading elements from a holey array. Depending on the CheckHoleMode,
they either return the hole as undefined or some NaN, or deoptimize
if the value is the hole or the hole NaN.

R=jarin@chromium.org
==========

to

==========
[turbofan] Introduce CheckHole and CheckHoleNaN operators.

These simplified operators are used to perform the hole checks when
loading elements from a holey array. Depending on the CheckHoleMode,
they either return the hole as undefined or some NaN, or deoptimize
if the value is the hole or the hole NaN.

R=jarin@chromium.org

Committed: https://crrev.com/502dd40c3307aaa60ee24251f13c9d36e11923a8
Cr-Commit-Position: refs/heads/master@{#37001}
==========

[commit-bot: I haz the power, 2016-06-15 12:25:39]

Patchset 2 (id:??) landed as
https://crrev.com/502dd40c3307aaa60ee24251f13c9d36e11923a8
Cr-Commit-Position: refs/heads/master@{#37001}